antifragile Archives - Algorithmic and Mechanical Forex Strategies

The guy that bet on Leicester City every year

September 5, 2016 by Shaun Overton Leave a Comment

Leicester City started the 2015 season with terrible odds of winning the Premier League Championship. Bookmakers only game them odds of 5,000:1 of winning.

To put that in context, you are more likely to die riding a bicycle than you were to win a bet on Leicester City. Or, you can think of betting on Leicester City every year. If you bet on them every single year for 5,000 years, you would expect them to win a grand total of… once.

2014 was hardly an indicator of their pending success. They were nearly relegated to a lower division (i.e., kicked out of the Premier League). And yet, they did win the championship last year.

Leicester City’s Biggest Fan

Meet John Michklethwait. He’s the former editor-in-cheif at The Economist and he’s currently editor-in-chief for Bloomberg. Clearly, he’s a very smart man. And yet, despite the odds and repeated disappointments, John bet on his old love, Leicester City, every single year dating back to the 1980s. That’s roughly 30 years of nonstop losing.

It wasn’t a lot of money each year: just £20. We all have our indulgences. I see the value of having skin in the game. £20 on a season is enough to make one care, but not so much that he’s upset about losing it.

Then something disruptive happened. John moved to the US last year for his position at Bloomberg. The chaos of the move threw him out of sorts, and he accidentally forgot to bet on Leicester City in 2015. He bet on them every single year dating back nearly 30 years. And yet the one year that he forgets to bet, not only did Leicester City win, but the bet paid out 5,000:1.

Let’s step back and calculate the cost of that oversight for Mr. Micklethwait.

£20 * 5,000 = £100,000.

A hundred… thousand… pounds. That kind of winning would put a nice dent in your mortgage, wouldn’t it?

The risk of low probability strategies

Everyone hears anecdotes about successful trend traders. Even winning only 30-40% of the time, they walk away big winners over time.

You live HERE. Math isn’t good enough. You also need to wonder if your strategy can handle real-world problems.

What if they took that even lower? They could move their stop losses closer to the market. They’d reduce the size of the average loser, but the winning percentage might also drop to 10-20%.

Mathematically, this could work out identically. 30% winners that earn 5x the average loser make for a profit factor of 1.5. A strategy with only 10% winners that make 15x the typical loser also have a 1.5 profit factor.

They’re the same. Aren’t they?

Planet Earth isn’t the same as planet Math. In the real world, people get sick and miss trades. Or, they move across the Atlantic and forget to place a £20 bet.

People move. They get sick. Computers break. Things can and will go wrong with trading.

Richard Dennis once commented that the Turtle Traders would often make their annual returns off of one, single trade. A single trade!

When your performance depends on positive outliers, you’re massively vulnerable to accidents. What happens if you’re sick that day? Or your internet goes down? Or your broker locks you out of your account on the worst possible day?

Life happens, brother. A plan that depends on perfection is no plan at all. You need to make yourself robust to failure. Or even better, you’d make yourself antifragile.

Winning percentages

I mentioned that you can do really well winning 30-40% of time. Why then, does my own trading strategy, Dominari, win 68% of the time?

Because I’m exploiting compound, exponential growth. It’s not just how much you win, but the order in which you win it.

Let’s take two examples:

A ranging strategy with a profit factor of 1.3 that wins 68% of the time.
A trending strategy with a profit factor of 1.3 that wins 30% of the time.

Look at the red circles. Trending strategies are prone to extreme outcomes, both positive and negative.

Each strategy risks about 1% on any given trade. And, the average of the range and trend strategies are identical in the long run.

But… and this is an important “but”, the expected worst case scenario with the trending strategy is substantially more likely compared to the range trading strategy. In effect, the average is more average with a ranging strategy than with a trending strategy.

Why is that? Because unusual losing streaks are devastating to trending strategies. Have you ever had a losing streak? It happens to everyone.

By using a strategy with a higher winning percentage, you’re making yourself robust to streaks of losers. And, not to mention, your average length of a winning streak is considerably higher.

Even though you’re getting the same mathematical outcome, you’re making things much easier on your trading psychology when you adopt a strategy with a higher winning percentage.

Dominari & Exponential Growth

You may have thought to yourself, “68%? That’s kind of a strange number to pick.”

You’d be right. The choice of 68% winners was not a coincidence. It is, in fact, the win rate on my Dominari strategy.

Dominari is about more than just buying and selling. Trading is also about managing a portfolio and position sizing. Position sizing is phenomenally important over your trading career.

My backtest results for Dominari show that for every $2,500, the account increased to $17,855.35 after 3 years. That kind of compound growth doesn’t happen by accident. That’s why I’d like to share the good news with you in my webinar this week.

I’m going to show you how to put that exponential awesomeness to work in your trading account. Sound good? Click here to register for the FREE webinar.

QB Pro Jan 15 Review

February 4, 2015 by Shaun Overton 2 Comments

January finished in the black, although the aftermath of the USDCHF chaos gave me a solid punch in the nose. It wasn’t the revaluation that got me. It was watching the chop and deciding that it was ok to turn the system back on.

Knowing what I knew at the time, the decision to turn the USDCHF trading back on was sensible and informed. The trouble came from the rumors of SNB intervention, which caused a 400 pip rally. Whether or not the SNB will institute a trading band for the franc against the euro is unclear.

I don’t want to get caught up in a market where the largest buyer’s goals are irrational. So… QB Pro is not currently trading francs.

Total performance for the month was 7.23%. I’m sorry to see that some of the new investors joined just in time to catch the worst trade in the system’s history. That obviously doesn’t feel very nice for them or myself. I’m working hard to reduce the risk of that happening again.

I’ve long mentioned my interest in reducing the 8.5% monthly blow-up risk on the current leverage. QB Pro 2.0 is in the works, a system that

Significantly reduces parameter sensitivity
Trades the systems as a portfolio rather than individual currencies

My goal is to turn QB Pro from a robust strategy into an antifragile strategy. Unfortunately, no charting package allows me to do portfolio level analysis in the way I wish. So.. I’ve spent the last two months writing a custom backtesting platform.

My whiteboard planning for a custom forex backtester

The initial results look excellent. But before I get too excited, I have one of my trusted developers reviewing the code to ensure that my out of sample tests are accurate. If the tests are indeed accurate, I hope to roll out the system live within the next 60 days. Stay tuned for updates to the system.

Via Negativa – Improvement Through Subtraction

July 5, 2013 by Andrew Selby Leave a Comment

Many traders get caught in the trap of constantly trying to add components to improve their systems. It’ often better to do the exact opposite.

Removing Negative Influences

“So knowledge grows by subtraction much more than by addition – given that what we know today might turn out to be wrong but what we know to be wrong cannot turn out to be right, at least not easily.” – Nassim Taleb

In Book VI of Antifragile, Nassim Taleb argues that removing negative influences can have a far greater affect on an outcome than adding positive influences. This is because it is far more likely for something we believe to be beneficial to prove harmful than for something we believe to be harmful to prove beneficial.

This claim is backed up by years of evidence. For example, in the ’50s (and as seen on Mad Men), the harmful effects of smoking were not widely acknowledged. People didn’t comprehend how terrible it was for their health, and some actually believed smoking had positive benefits. It is much harder to find examples of things that were believed to be negative that were later proven to be beneficial.

via negativa encourages removal of complexity

There was a time where people actually believed that smoking was healthy.

The same concept can be found in trading. For decades now, financial analysts have been touting “the power of compounding interest” from buy-and-hold strategies. However, with every black swan event that leads to severe crashes, more and more investors are seeing the error of that logic. There are far fewer examples of strategies that were commonly believed to be flawed and later proved profitable. Can you think of any?

The Future is in the Past

“Effectively my answer would be to make them read the classics. The future is in the past.” – Nassim Taleb

According to Taleb, there is a direct correlation between the length of time an idea has existed and its value to someone studying it. Applying that concept to system trading, we should focus our time on systems that have a proven track record over decades of market performance.

One of the great things about trading systematic strategies is that we can backtest those strategies to see how they would have performed historically. While there are plenty of biases that should be considered with these results, past data can still give us an excellent idea of how these strategies might perform in the future.

A system that has performed well over the past century has a much better chance of performing well over the next century. A system that performed poorly over the past century is far more likely to perform poorly over the next century.

Smoothing Volatility & Increasing Fragility

“what we call diseases of civilization result from the attempt by humans to make life comfortable for ourselves against our own interest, since the comfortable is what fragilizes.” – Nassim Taleb

As humans, we have a natural motivation to attempt to smooth out the volatility of our trading systems. High volatility systems have the ability to keep us up at night worrying about their risk of ruin.

While attempting to smooth out some volatility so that you can get a good night’s rest is important, it can often be taken too far. Many times, traders add too many volatility filters and create overly curve-fit systems that become highly vulnerable to Black Swan events. Traders need to remember that a comfortable system with low volatility is often the most fragile type of system.

Quitting Smoking & Cutting Losses

“telling people not to smoke seems to be the greatest medical contribution of the last sixty years.” – Nassim Taleb

The simple realization that smoking was terrible for long-term health may have saved more lives than every other “innovation” in that time combined. Likewise, the simple realization that taking critical losses is terrible for your portfolio’s long-term survival may be more important than any technical indicator invented in the last sixty years.

Focusing on all of the newest derivative pricing models will not do anything to help your trading if you miss the key point of cutting your losses. Therefore, Via Negativa suggests that you should cut out all those new, innovative approaches and focus on what you know for a fact will hurt you, losing money.

People Want To Be Fragile

June 3, 2013 by Andrew Selby Leave a Comment

While reading through Nassim Taleb’s applications of convexity, the planning fallacy, and Jensen’s inequality, I began to wonder why more people don’t take these concepts and actually put them to use. Why are people content to employ fragile trading strategies and on a larger scale, fragile life strategies?

Taleb introduces the concepts of linear vs nonlinear and convexity vs concavity in Book V of Antifragile. Some things, the fragile, break under stress. There are other things, the antifragile, that perform better under stress.

People constantly set themselves up to be brutally hurt by Black Swan events. It’s almost like these people are choosing to be fragile.

Why Don’t People Want To Become Antifragile?

If anyone in the world can order Taleb’s book from Amazon for less than $20, why aren’t more people interested in protecting themselves from future Black Swan events by becoming more antifragile? Even if they can’t afford the price of the book, there is an amazing amount of free information available on the internet. All anyone needs to access it is a free library card and a little bit of effort.

Ed Seykota famously said that “everyone gets what they really want out of the market.” His argument is that people ignore all of the available evidence that their trading system is flawed because they subconsciously want to fail. Is it then possible that people ignore antifragility because they want to be fragile?

This idea has a broader application than just trading systems. In all aspects of life, people are perfectly content to live in the bliss of their own ignorance.

Stockholm Syndrome

In a Facebook post this weekend, Michael Covel suggested that the definition for Stockholm Syndrome could be applied to buy and hold investors:

Fragile people seek things other than profits

You ultimately get what you want out of the markets

This same thinking can be applied to Taleb’s fragility. When presented with all of the evidence documenting how fragile their investments are, most people will still passionately defend their strategies. It is almost as if they have been brainwashed into believing that they deserve to suffer from the occasional Black Swan event.

Taleb takes this argument even deeper when he discusses fragility with respect to randomness.

Better Than Random

“The hidden harm of fragility is that you need to be much, much better than random in your prediction and know where you are going, just to offset the negative effect.” – Taleb

Taleb points out that in order to successfully trade a fragile strategy, you have to perform dramatically better than random. You have to have a tremendous edge. You have to be brilliant.

While some people have proven that they do have such an edge, the majority of traders post results that are far worse than a completely random strategy. This leaves them completely vulnerable to the next Black Swan event. Yet they continue to trade this way.

Worse Than Random

“The hidden benefit of antifragility is that you can guess worse than random and still end up outperforming.” – Taleb

This is the strength of many systematic trend following strategies. In recent weeks, I have looked at the 83/17 Breakout System and the 10/100 SMA Long Only System. These two systems are profitable on 41.46% and 42.53% of their trades respectively, yet both systems are profitable overall.

The positive risk to reward ratio built into each system gives its returns the convexity that Taleb discusses. These systems are antifragile.

Capitalizing On The Trend

There are antifragile trading systems that are widely available for free on the internet, but people don’t take advantage of this. There are also more antifragile lifestyles available, but most people prefer to stick to the normal, fragile world they know that involves a 401k and keeping up with the Joneses.

While solving this global sociology problem is a little bigger than the scope of this post, our trading can benefit from recognizing its existence. Since trading is a zero sum game, our profits are someone else’s losses. By being aware of the fragile strategies most people employ, we can position ourselves to take advantage of the same Black Swan events that will ruin our fragile counterparts.

Less Is More: The Halo Effect & Green Lumber

May 23, 2013 by Andrew Selby Leave a Comment

It is often said that system traders spend too much time attempting to perfect entry and exit points. What if we expand that idea and take a look at whether or not we are putting too much effort into all of our trading efforts?

There is more to be learned from actually trading than is taught in classrooms. You can trade successfully right now, despite the fact that there are thousands of people who have more knowledge than you.

When Two Things Are Not “The Same”

In Book IV of Antifragile, Nassim Taleb spends a good portion of the chapters addressing the lack of actionable knowledge that can be gained from institutional learning. He strongly advocates a self-directed style of learning as opposed to the standard university approach.

As systematic traders, we are forced to follow the self-directed approach that Taleb mentions because systematic trend following strategies are not taught in expensive universities.

“When you are fragile you need to know a lot more than when you are antifragile. Conversely, when you think you know more than you do, you are fragile (to error).” – Taleb

Taleb uses both the Halo Effect and The Green Lumber Fallacy to illustrate his point that knowledge about a topic does not necessarily lead to success in that field. In trading, it is popular to discuss why markets behave in different ways.

This need to understand why something happens increases fragility. In order to make our trading systems more antifragile, we must focus less on what we know and more on what we don’t know. We don’t know the future.

Halo Effect

“The more interesting their conversation, the more cultured they are, the more they will be trapped into thinking that they are effective at what they are doing in real business (something psychologists call the halo effect, the mistake of thinking that skills in, say skiing translate unfailingly into skills in managing a pottery workshop or a bank department, or that a good chess player would be a good strategist in real life). – Taleb

The most commonly used example of the halo effect, which was first researched by Edward Thorndike, occurs when our opinion of someone’s character traits are influenced by their attractiveness.

The halo effect applies well to brand marketing. The amazing popularity of Apple’s iPod has resulted in great success for many of its other products. I personally have experienced this effect with Apple products.

After being a long time iPod user, I purchased an iPad a few years ago and was completely blown away. My enthusiasm for the iPad convinced me that I next needed to acquire an iPhone. Six months later, I was the proud owner of an iPhone 4S, which has a woman inside of it that talks to me. Now that I had an iPod, iPad, and iPhone, the decision on what kind of laptop to buy was already made for me. I wanted a MacBook Air.

While the halo effect worked out nicely for Apple, Taleb suggests that it increases fragility. In the last part of the quote above, he compares the skill of playing chess with real life strategy. The obvious connection to trading is that knowledge about trading does not translate into profitable trading.

An extensive working knowledge of the Black Scholes Model will make you sound intelligent at a dinner party, but there is a lot more that is needed to actually trade options successfully. A trader with far less knowledge can benefit simply by not over thinking his trades.

He is free to experiment with trial and error the way Taleb suggests. The learn as you go approach is far more antifragile. This is how the Green Lumber Trader operates.

Green Lumber

You don’t have to know everything about the lumber industry to trade it.

The Green Lumber Fallacy is a reference Taleb makes to a commodity trader who traded “green lumber” successfully, even though he didn’t know why it was called that. The theory behind this example is that the information that is generally associated with a market may have little to no bearing on that market. Taleb also gives an example of a Swiss franc trader who “could not place Switzerland on the map.”

“So let us call the green lumber fallacy the situation in which one mistakes a source of necessary knowledge – the greenness of lumber – for another, less visible from the outside, less tractable, less narratable.” – Taleb

Again here, we see success coming more from the lack of knowledge than an abundance of it. This idea reminds us that many of the issues and strategies discussed on television simply have no bearing on our trading results. Rather than drowning ourselves in data, the correct way to build a successful system is to look at less data. Less truly is more.

Making Trading Systems More Antifragile

May 13, 2013 by Andrew Selby Leave a Comment

In order to make trading systems more antifragile, we must focus on changing their risk exposure. If your primary concern in trading is your system’s profitability, then you are setting yourself up for an epic failure. The first and foremost concern of any trading system absolutely must be the risk within a system.

Risk of Ruin

In Book III of Antifragile, Nassim Taleb presents the argument that profit is a secondary goal in business. He argues that anyone who primarily focuses on profit is underestimating their risk exposure.

This concept applies directly to trading systems. Far too many traders use potential returns as the measuring stick for a trading system. Following Taleb’s argument, traders should put far more emphasis on the nature of the system.

“This fragility that comes from path dependence is often ignored by businessmen who, trained in static thinking, tend to believe that generating profits is their principal mission, with survival and risk control something to perhaps consider – they miss the strong logical precedence of survival over success.” – Taleb

While generating profits is certainly the end goal of any system, the ability to generate those profits requires the survival of the system. If the system goes bust at any point, then those future profits mean nothing.

Even if the system generates huge profits for years, if there are significant risks that could eventually catch up to it, then the system is just trading on borrowed time. The most important criteria for assessing a system’s fragility is its risk of ruin.

Risk and reward should be balanced against the long term objective of continuing to trade.

The least desirable way to protect against risk of ruin is to decrease the fragility of the system. This may sound like common sense, but far too many traders never make the connection.

As I covered earlier this week, there are three way of doing this: increasing the system’s win rate, increasing the system’s profitability, or decreasing trading size. Of these options, the only one that is completely in the trader’s control is the amount of risk their capital is exposed to.

Sensitivity to failure is best analyzed by adjusting the accuracy percentage and payout ratios. If you trade a trending system and the average payout drops by 20%, what happens to the system? Many trending systems completely fall apart.

Sometimes, a difference of 10% between the historical returns and future observations makes the difference between decent profits and huge losses. This is far more true when the expected percent accuracy falls and the payoff falls, too. Minor errors in observation lead to drastically different outcomes.

Fragility is Relative to Upside/Downside Risk

In the preceding chapter, Taleb explains that fragile systems have more to lose than they can gain. The opposite is true for antifragile systems. They have more to gain than to lose. They have a limited downside.

“You are antifragile for a source of volatility if potential gains exceed potential losses (and vice versa).” – Taleb

This further supports the argument that risk of ruin should be the primary concern when evaluating a trading system. Regardless of the system’s apparent upside, if the downside is too great, the system is no good. If the potential exists to lose everything, there is no possible upside that can balance that.

Long Term Capital Management

Taleb uses the example of a gambler to illustrate that no matter how good his strategy is, if he is exposed to the risk of losing everything then nothing else matters. The same can be said of anyone who’s trading system exposes them to too much risk.

It doesn’t matter how efficient the strategy is or how impressive the returns are if the risk of losing everything is too great. Risk of ruin trumps all other factors because it has the ability to end the game.

Long Term Capital Management (LTCM) provided an excellent example of this concept in the late 1990s. The fund was believed to have superior strategies implemented by brilliant traders and many investors expected the strong returns to continue indefinitely. The problem is that no one was concerned with the downside risk, leaving the fund vulnerable to a Black Swan event, which happened in 1998.

The combination of some unexpected events with a big appetite for leverage proved to be disastrous for LTCM and its investors. LTCM believed that it had smoothed out volatility and could provide consistent returns. As we all learned, the short-termist obsession with pretty Sharpe ratios on high leverage came at the cost of making a blowup inevitable..

Antifragile Trading – Enhanced By Volatility

April 23, 2013 by Andrew Selby 2 Comments

In Antifragile, author Nassim Nicholas Taleb presents the argument that the opposite of fragile is not simply something that is unaffected by stress or pressure. He argues that there is a quality called antifragile where something actually improves as a result of that stress or pressure. Taleb categorizes things on a scale from fragile to antifragile, with a mid-point he calls robust.

When this idea is applied to trading, the first comparison that jumps out at us is between buy and hold investors and systematic trend following traders. Systematic trend following traders excel under extremely volatile situations, are virtually ignored by intellectuals in their field and are capable of handling deviations because they are not predictive in nature.

An Illustration of The Antifragile Concept

Taleb illustrates his concepts of fragile, robust, and antifragile through Greek and Roman mythology.

Fragile is represented by the story of Damocles, who was invited to a dinner at which there was a sword hanging over his head that was held up by a horse’s hair. Damocles is considered fragile because at any second the hair could snap dropping the sword onto him.

The concept of robust is represented by the Phoenix, which is a bird that is able to be reborn from its own ashes. No matter how many times you attempt to destroy the Phoenix, it continues to return to its original state, unharmed.

Antifragile is represented by the Greek Hydra, which had multiple heads. Every time one of its heads was chopped off, two more would grow back. Any attempt to hard this creature actually made it grow stronger. This is the core concept of antifragile.

An Application To Trading

One of the characteristics that can help measure the fragility of something is how it responds to volatility. This allows us to easily apply this concept to trading, because volatility is actually measured in most markets.

VIX Monthly shows antifragile periods of strength

A monthly VIX chart dating back to 1985

If we look at a long term chart of the VIX, we can see four distinct areas of extreme volatility: The Market Crash in 1987, The Asian Currency Crisis in 1997-98, The Dot-Com Bubble & Crash from 1998-2002, and the Financial Collapse in 2008. These highly volatile periods were high stress situations that would have made antifragile trading strategies better while making fragile strategies worse.

It is no coincidence that it was during these exact periods that systematic trend following traders posted some of their best years, and at the same time, buy and hold investors lost fortunes.

Systematic trend following strategies are not only built to handle these kinds of high stress, high volatility situations. They become stronger from them simply because they provide traders with a way to change direction. This ability to go with the flow during even the most chaotic times is a stark contrast to the rigidity of a buy and hold strategy where the investor has no choice but to sit and watch their portfolio crumble.

Overcompensation & Overreaction

“Intellectuals tend to focus on negative responses from randomness (fragility) rather than the positive ones (antifragility).” – Taleb

Again here, we can see a clear application of this with respect to trading during high volatility periods. In each of our four high volatility trading periods from the past 25 years, most of the intellectuals and so-called market experts were focused on the negative responses, which were the buy and hold investors who were losing their fortunes left and right.

During each of these periods, there was very little focus on traders who were benefiting from the increased volatility. The antifragile traders seemed to go unnoticed, and we know now that systematic trend following traders where the ones who were benefiting from these volatile markets.

Layers & Deviation

Taleb also discusses the complexities that can arise when looking at fragility through different layers and dimensions. He illustrates that since what is good for the antifragile is bad for the fragile, then what is good for the fragile is likely bad for the antifragile.

We have already discussed the huge success that antifragile trend following strategies have seen during times that were terrible for fragile buy and hold strategies. When the inverse situations occurs where markets have little volatility, we have seen that systematic trend followers and buy and hold investors can produce rather similar results. The key difference though, is that the antifragile systematic trend followers have the ability to adjust when a black swan shows up.

“When you are fragile, you depend on things following the exact planned course, with as little deviation as possible – for deviations are more harmful than helpful.” – Taleb

He continues to explain that the fragile is very predictive and that being so predictive is exactly what causes fragility. This is the problem with buying and holding with the expectation that a position will go up over time. The entire strategy is predictive in nature, and that is exactly why it becomes fragile. It can’t handle a deviation from what it predicts.

On the contrary, a systematic trend following approach can, by definition, handle any deviation from what is expected because it simply goes along with that deviation. This ability to adapt and change is magnified during periods of high volatility, which is when systematic trend following is at its best.